Strong Coupling between a Trapped Single Atom and an All-Fiber Cavity

We demonstrate an all-fiber cavity quantum electrodynamics system with a trapped single atom in the strong coupling regime. We use a nanofiber Fabry-Perot cavity, that is, an optical nanofiber sandwiched by two fiber-Bragg-grating mirrors. Measurements of the cavity transmission spectrum with a single atom in a state-insensitive nanofiber trap clearly reveal the vacuum Rabi splitting.

Figure 1

(a) Simple schematic of the experiment. Two FBG mirrors form a Fabry-Perot cavity with a nanofiber waist. Two-color trapping beams (${\lambda }_{\mathrm{blue}},{\lambda }_{\mathrm{red}}$) create a nanofiber trap for an atom. A probe pulse with a frequency ${\omega }_P$ is used to detect the presence of an atom and to measure the vacuum Rabi spectra. DM, dichroic mirror; $F$, filters; APD, avalanche photodetector. The inset shows the profiles of the nanofiber and tapered regions measured by a scanning electron microscope (blue circles) and a numerical model (green solid line). (b),(c) Transmission spectra separately measured for FBG1 and FBG2, respectively. The red solid line indicates the atomic resonance frequency, ${\omega }_A$ ($6^2{S}_{1/2};F=4\to 6^2{P}_{3/2};F^{\prime }=5^{\prime }$ transition of cesium). The inset in (c) is a magnified plot for the region around ${\omega }_A$. (d) Trap potential in the vicinity of the nanofiber surface. The trap frequencies in the axial, radial, and azimuthal directions are $2\pi \times (267,159,36)\mathrm{kHz}$, respectively. Details of the trap potential structure can be found in Ref. [33].

Figure 2

(a) Histogram of the transmission intensity of the detection probe with (blue) and without (gray) the optical molasses. The dark count of the detector is $1.4\times {10}^3$ counts per second (cps), and the measurement background noise including the dark count is about 10 kcps. (b) Transmission spectra as functions of the probe detuning $\mathrm{\Delta }$. Data sets and fits are normalized to the empty-cavity transmission and are vertically offset for clarity. The observed asymmetry in the spectra is presumably due to the effect of the probe pulse on the center-of-mass motion of the atom. Error bars are the standard error of the mean. The atom-cavity coupling rates $g$ for the fits are $2\pi \times (1.3,1.9,2.9,4.3,7.8)\mathrm{MHz}$ for (ii)–(vi), respectively.

Figure 3

(a)–(d) Vacuum Rabi splitting spectra with various atom-loading times ${\tau }_{\mathrm{load}}=20$, 10, 5, and 2 ms, respectively. The atom-cavity coupling rate $g$ determined from the fits are $2\pi \times (7.7\pm 1.1,8.9\pm 1.4,7.0\pm 1.1,7.8\pm 1.1)\mathrm{MHz}$ for (a)–(d), respectively. The probability of the corresponding events $P_{(\mathrm{vi})}$ is shown in each panel [35].

Figure 4

Normalized transmission as a function of the hold time. The solid line is an exponential fit to the data. Error bars denote the standard error of the mean.